US20080273635A1

US20080273635A1 - Frequency-offset estimation

Info

Publication number: US20080273635A1
Application number: US11/743,983
Authority: US
Inventors: Holger Neuhaus; Bertram Gunzelmann; Michael Weber
Original assignee: Individual
Current assignee: Infineon Technologies AG; Intel Corp
Priority date: 2007-05-03
Filing date: 2007-05-03
Publication date: 2008-11-06
Also published as: US7899136B2; DE102008016434A1; DE102008016434B4; DE102008064822B3

Abstract

A device for calculating an overall frequency-offset estimate from a plurality of frequency-offset estimates includes a receiver unit configured to receive a plurality of data streams, a phase shift estimation unit configured to generate a phase shift quantity representative of a phase shift over a number of consecutive data samples in the respective data stream, and a phase shift processing unit coupled to the phase shift estimation unit and configured to calculate a quantity related to the signal-to-noise ratio associated with a data stream on the basis of the respective phase shift quantity. Further, the device includes a frequency-offset estimation unit configured to generate frequency-offset estimates associated with the respective data streams, and a combiner configured to calculate an overall frequency-offset estimate on the basis of the frequency-offset estimates and on the basis of the quantities related to the signal-to-noise ratios.

Description

FIELD OF THE INVENTION

The invention relates to the estimation of an overall frequency-offset on the basis of a plurality of frequency-offset estimates and further to the estimation of a quantity representative of a signal-to-noise ratio. In particular, the invention relates to communications systems and receiver.

BACKGROUND OF THE INVENTION

In mobile communication systems a mobile station may be connected to multiple base stations at the same time. In this case, frequency-offset correction in the mobile station is typically not done separately for each base station. Instead, frequency offset correction aims at correcting offsets of the frequency generating device. Usually, only a single device is used as a reference for frequency generation. Therefore, if frequency-offset estimates are obtained by measurement from a plurality of signal sources such as, for example, base stations to which the mobile station is connected, a strategy has to be devised of how to combine these multiple frequency-offset estimates into a single value, i.e. an overall frequency-offset estimate. As the reception performance and transmit frequency offset on the uplink of a mobile station sensitively depends on the compensation of frequency-offsets, an efficient strategy for combining the frequency-offset estimates into a single value is desirable.
Further, quantities representative of signal-to-noise ratios of signals in a radio receiver are known to be necessary for many applications and are used in a radio receiver for a variety of computations. Therefore, efficient algorithms and circuitry to generate signal-to-noise estimates in a radio receiver are desirable.

BRIEF DESCRIPTION OF THE DRAWINGS

Aspects of the invention are made more evident by way of example in the following detailed description of embodiments when read in conjunction with the attached drawing figures, wherein

FIG. 1 is a block diagram showing basic functional blocks of a receiver;

FIG. 2 is a block diagram of the automatic frequency control unit shown in FIG. 1;

FIG. 3 is a schematic illustration of an embodiment of the combiner shown in FIG. 2;

FIG. 4 is a schematic diagram of a device for combining multiple frequency-offset estimates in a receiver; and

FIG. 5 is a schematic diagram of a device for calculating signal-to-noise ratio estimates.

DETAILED DESCRIPTION OF THE INVENTION

In the following description further aspects and embodiments of the present invention are disclosed. In addition, reference is made to the accompanying drawings, which form a part thereof, and in which is shown by way of illustration, one or more examples in which the invention may be practiced. The embodiments herein provide a better understanding of one or more aspects of the present invention. This disclosure of the invention is not intended to limit the features or key-elements of the invention to a specific embodiment. Rather, the different elements, aspects and features disclosed in the embodiments can be combined in different ways by a person skilled in the art to achieve one or more advantages of the present invention. It is to be understood that other embodiments may be utilized and structural or logical changes may be made without departing from the scope of the present invention. The elements of the drawing are not necessarily to scale relative to each other. Like reference numerals designate corresponding similar parts.
At a receiver, the actual frequency of the received signal may, for various reasons, be different from a local frequency used in the receiver to process the received signal, thereby causing a frequency-offset in the receive signal path of the mobile receiver. For instance, frequency shifts in the received radio signal or detuning of the local oscillator of the receiver may account for the generation of such frequency-offset in the receiver. A frequency-offset in the receiver significantly impairs the performance of the receiver and may lead to a transmit center-frequency offset that exceeds allowed tolerance levels.
In the following, according to a first embodiment, an improved combining of frequency-offset estimates in a receiver is described. The embodiments described below may be employed in various mobile communications systems, among them CDMA (Code Division Multiple Access) systems such as, for example, UMTS (Universal Mobile Telecommunications System) or mobile communications systems using other types of multiple access schemes, for example, GSM (Global System for Mobile Communications). Communications systems considered below may be, for example, multi-cell systems or single cell systems comprising multiple transmitters.
FIG. 1 illustrates basic functional blocks of a receiver, for example, a mobile receiver. The receive signal path of the receiver may comprise an amplification stage 1, a first reception section (RX1) 2, a received signal strength indicator (RSSI) unit 3, an automatic gain control (AGC) unit 5, a second reception section (RX2) 4, an automatic frequency control (AFC) unit 6 and a data processing unit (DATA) 7.
Amplification stage 1, first reception section 2, RSSI unit 3 and AGC unit 5 form a feedback loop used for received signal power control. More specifically, amplification stage 1 receives at input 10 an input signal provided by a receiver input. The receiver input may be an antenna (not shown) followed by optional circuitry such as filters and/or additional signal preprocessing units. The input signal is amplified (e.g., multiplied) by a power control signal connected to the second input 11 of the amplification stage 1. An output of the amplification stage 1 is fed into the first reception section 2 of the receiver. The first reception section 2 typically comprises the RF (Radio Frequency) part of the receiver. Thus, as known in the art, the input signal is down-converted to an intermediate frequency (IF) band or the baseband. Down-conversion is typically done by a mixer (not shown) which is operated by a local oscillator 8 outputting an oscillator signal of frequency f_os. Further, the first reception section 2 may comprise filter stages and other signal processing units as known in the art.
An output signal of the first reception section 2 is fed into the RSSI unit 3. The RSSI unit 3 generates an output 12 which is representative of the signal strength or signal power of the received signal. The output 12 of the RSSI unit 3 is coupled to an input of the AGC unit 5. The AGC unit 5 compares the output 12 of the RSSI unit 3 to a target value and generates an amplification control signal 13 in response to the comparison result. The amplification control signal 13 is fed into a control input 11 of the amplification stage 1 and thus is used to control the received signal power amplification of the receiver. In other words, the feedback loop comprising the amplification stage 1, the first reception section 2, the RSSI unit 3 and the AGC unit 5 establish a regulation circuit which is operative to adjust the signal power in the received signal data stream at the signal output 14 of the RSSI unit 3 to a fixed, known energy target level Pt′.
The signal output 14 of the RSSI unit 3 is fed into the second reception section 4. The second reception section 4 of the receiver may comprise a demodulator for reconstructing data samples (also termed symbols in the following) contained in the received signal data stream. To this end, the second reception section 4 may for instance comprise parts of a RAKE receiver for detection and preprocessing of signal contributions received over different propagation paths of a radio channel. Further, the second reception section 4 may be equipped with a descrambling and a despreading stage used for extracting the wanted data and control/pilot signals from the received signal which is a superposition of all data and control signals picked up at the input. As known in the art, in CDMA communications systems, user signals (i.e. logical channels per base station and user) are separated by different spreading codes. Thus, user signal separation in the receiver may be performed by despreading the received signal.
Further, the second reception section 4 provides for separation of signals transmitted from different signal sources such as base stations. In CDMA communications systems, each signal source uses an individual scrambling code for coding signals to be transmitted. Therefore, separation of signals received from different signal sources in the receiver may be accomplished by descrambling the received signal.
In communications systems using other techniques for multi-user access and/or base station identification, it is likewise possible to distinguish between signals which are intended for different users and/or received from different signal sources (e.g., base stations), respectively. In other words, at an output 15 of the second reception section 4, a user-specific and signal source-specific signal may be provided. Such signal may be represented by a stream of complex-valued symbols r_i,massociated with the particular mobile station, where i denotes the time index and m identifies the base station, m=1, . . . , M, from which the symbol r_i,mhas been transmitted.
The output 15 of the second reception section 4 is coupled to an input 16 of the AFC unit 6 and an input 17 of the data processing unit 7. Data processing in the data processing unit 7 may be accomplished according to known techniques in the art (e.g., may comprise de-interleaving, channel-decoding etc.) and will not be described in more detail in the following. In the AFC unit 6, the symbols r_i,mare processed to calculate an overall frequency-offset estimate Δf. The overall frequency-offset estimate Δf is used as input for a control unit 9 which generates a frequency control signal f_cfor adjusting the frequency f_osgenerated by, for example, an oscillator 8. Thus, the first reception section 2, RSSI unit 3, second reception section 4, AFC unit 6, control unit 9 and oscillator 8 form a frequency closed-loop control circuit which is operable to maintain ΔA=0. The control unit 9 may be implemented by a look-up table or a conversion function and affects the control response of the frequency closed-loop control circuit, in one embodiment. In more general terms, the overall frequency-offset estimate Δf may be used to control the processing of the received signal in the mobile receiver.
Various modifications are feasible for implementing the frequency closed-loop control circuit according to the invention. For instance, the frequency control signal f_cmay be input to a frequency correction stage (not shown) instead of oscillator 8. The frequency correction stage may be arranged at any point in the receive signal path between the antenna and the input 17 of the data processing unit 7. Such frequency correction stage may either be located in the analog part or in the digital part of the receiver. In the latter case, frequency correction may for instance be performed in the second reception section 4 by means of, for example, a digital frequency correction circuit (not shown). In this regard, it is to be noted that analog-to-digital conversion is typically performed in the first reception section 2 of the receiver but may generally be performed at any convenient point in the receive signal path shown in FIG. 1.
FIG. 2 illustrates by way of example the design of the AFC unit 6 according to one embodiment. The AFC unit 6 comprises a phasor generation unit 100, an averaging unit (AVG) 101, a unit 102 computing the squared magnitude of the complex phasors, a divider (DIV) 103, a combiner 104 and an argument unit 105. The phasor generation unit 100 computes on the basis of the input symbols r_i,mcomplex-valued phasors p_i,m. Phasor p_i,mmay be defined in one embodiment by
p _i,m =r _i,m ·r _i−D,m, (1)
and is the product of symbol r_i,mand the conjugate complex of a preceding symbol r_i−D,m(associated with the same signal source m). D is an integer and may typically be chosen to be D=1. The argument of phasor p_i,m, i.e. arg(p_i,m), is indicative of a phase shift between actual symbol r_i,mand preceding symbol r_i−D,m.
Phasors p_i,mare fed into the averaging unit 101. In the averaging unit, phasors p_i,mare averaged or summed over an averaging length L_avgaccording to
$\begin{matrix} p_{m}^{av} = \sum_{i}^{L_{avg}} p_{i, m} . & (2) \end{matrix}$
The averaged phasors p_m ^avmay be used twofold, namely for computing frequency-offsets Δf_massociated with the symbol streams r_i,mand for computing quantities representative of a signal-to-noise ratio associated with the symbol streams r_i,m.
To generate an estimate of the frequency-offset Δf_m, the averaged phasors p_m ^avare fed into the argument unit 105. Argument unit 105 calculates the argument—i.e. the phase or angle—of each averaged phasor p_m ^av, which is proportional to the frequency-offset Δf_maccording to Δf_m=arg(p_m ^av)/(2π·L_avg·T_sym). Note that normalizing by the factor (2π·L_avg·T_sym)⁻¹is not depicted in FIG. 2. Further note that for D≠1, D·T_symhas to be used instead of T_sym.
To generate quantities representative of the signal-to-noise ratios, the averaged phasors p_m ^avare fed into the square unit 102 followed by the divider 103. At the output of the divider 103, a quantity C_mwhich relates to the signal-to-noise ratio associated with the symbol stream r_i,mis output.
In the following, it is shown that a quantity related to the signal-to-noise ratio associated with the symbol stream r_i,mmay be computed on the basis of phasors p_i,m. A quantity related to the signal-to-noise ratio may be any quantity which is related to the signal-to-noise ratio as a meaningful metric, particularly a quantity which translates into the signal-to-noise ratio by a unique functional relationship. To simplify notation, the index m is dropped because the computation is identical for symbols from each signal source (e.g. base station) m.
The energy of the averaged phasor p^avis the square of the absolute value of the complex-valued averaged phasor p^avand can be written as the sum of the square of the averaged phasor's real and imaginary part, i.e.
Q=|p ^av|² =re(p ^av)² +im(p ^av)². (3)
This energy relates to the signal-to-noise ratio of the signal. It is assumed that the receiver uses ideal automatic gain control based on RSSI measurements on chip level before base station separation and user separation (e.g., descrambling and despreading). To simplify matters, it is further assumed that all non-wanted signal parts contribute either orthogonally (e.g., are cancelled by despreading) or appear as white Gaussian random noise (such as, for example, signals from base stations on different scrambling codes). The following derivation uses the special case of a CDMA receiver.
The following notation is used below:

- E The wanted-signal energy per chip
- N The noise energy per chip
- SNR=E/N The signal-to-noise ratio on chip-level
- L_corrSpreading factor (number of chips per symbol)
- L_avgAveraging length for phasors
- Δf frequency-offset
- T_symSymbol sampling time
- Pt′ AGC energy target level per chip at the output of the AGC feedback loop

The signal energy at the output of the AGC feedback loop is the AGC energy target level, i.e.
Pt′=E+N (4)
with
N=Pt′/(1+SNR)
E=Pt′−N (5)
A frequency-offset of Δf reduces the average energy per chip E after despreading over L_corrvalues (corresponding to T_sym) to a “degraded” average energy per chip E_degaccording to
E _deg *L _corr ² =E*|sinc(Δf*T _sym)|² *L _corr ² (6)
For the quality indication of frequency-offset estimates, a degraded signal-to-noise ratio can be defined by
SNR _deg =E _deg /N with SNR _deg <SNR (7)
The squared averaged phasor energy Q can finally be expressed by
$\begin{matrix} \begin{matrix} Q = 4 (L^{2} E_{\deg}^{} + 2 {LNE}_{\deg} + L N^{2}) \\ = 4 {(Pt / (1 + SNR))}^{2} L^{2} {SNR}_{\deg}^{} (\begin{matrix} 1 + 2 / (L {SNR}_{\deg}) + \\ 1 / (L {SNR}_{\deg}^{2}) \end{matrix}) \end{matrix} with L = L_{corr}^{} L_{abg} . & (8) \end{matrix}$
Pt takes into account any possible power scaling by the second reception section 4, i.e. may be interpreted as the energy target level per chip downstream of the second reception section 4. If the second reception section 4 does not introduce any power scaling, Pt=Pt′.
A first approximation of this term can be derived on the basis of the following assumptions:
SNR<<1 (
SNR_deg<<1)
L*SNR_deg>>2.
Under these assumptions, equation (8) yields
Q=4Pt ²(SNR _deg ² L ² +L). (9)
A second even simpler approximation is found on the basis of more stringent assumptions:
SNR<<1 (
SNR_deg<<1)
L*SNR_deg>>2
L*SNR_deg ²>1
Under these assumptions, equation (8) yields
Q=4Pt ² L ² SNR _deg ². (10)
As it is apparent from equation (10), the energy Q of the averaged phasor is proportional to the square of the degraded signal-to-noise ratio SNR_degif L is large enough. As an example, in UMTS, when using the common pilot channel symbols as input symbol stream r_i,mfrom each base station m, the spreading factor has a fixed value of L_corr=256. Thus, L_avgmay be chosen such that equation (10) is roughly satisfied (note that in many cases not necessarily exact results are required). Using for instance an average length of L_avg=150 corresponding to one UMTS frame, the approximation set out in equation (10) holds for about −34 dB<SNR_deg<−10 dB. Translated from chip-level to symbol-level, the range of the degraded signal-to-noise ratio SNR_degis from −10 dB to +14 dB. This range should be fully sufficient to provide differentiation between different signal-to-noise ratios of frequency-offset estimates.
The averaging length L_avgmay be variable and be chosen to be different for different measurements. For this reason and to further limit the dynamic range of the results, the energy Q of the averaged phasor may be divided by L². This is done in the divider 103 according to
C=Q/L ². (11)
As the AGC energy target level Pt′ and thus the scaled energy target level Pt are likely to be fixed values, C is then a quantity representative of the signal-to-noise ratio of the processed signal irrespective of the used average length L_avg.
FIG. 3 illustrates an exemplary embodiment of the combiner 104. Combiner 104 may comprise an optional discarding unit (DISC) 200, a weighting unit (WG) 210 and an averaging unit (AVG) 220.
The discarding unit 200 is operable to discard frequency estimates Δf_mhaving a degraded signal-to-noise ratio SNR_deglower than a required minimum SNR_deg,min. To this end, the discarding unit 200 may comprise a multiplier 201, a comparator 202 and a multiplexer 203. The multiplier 201 is configured to compute a threshold value T according to
T=4 SNR_deg,min ²Pt² (12)
in one embodiment, wherein C_mis compared to threshold value T in comparator 202. If C_m<T, the corresponding frequency-offset Δf_mhas a low confidence because it has been calculated on the basis of noisy data with a degraded signal-to-noise ratio SNR_deg(which is represented by C_m) lower than the minimum degraded signal-to-noise ration SNR_deg,min. In this case, C_mis set to C_m=0 by multiplexer 203. It is to be noted that threshold values T different from the threshold value T indicated in equation (12) may be used.
Values of C_mas output by the discarding unit 200 are input into the weighting unit 210. The weighting unit 210 may comprise a first multiplier 211 and a second multiplier 212 connected in series. The (optional) first multiplier 211 multiplies each value C_mby a weighting factor J_m, in one embodiment. Weighting factor J_mmay provide for a source-importance weighting adjustment which introduces priorities with respect to the signal content. More specifically, a symbol stream r_i,mof one specific signal source (e.g. base station) m may be of higher importance for signal demodulation in the receiver because, for instance, base station m offers a specific service. As an example, base station m may provide HSDPA (High Speed Downlink Packet Access) in an UMTS communications system. Then, it may be desirable to optimize the frequency adjustment in the receiver such that the compensation of Δf_m is favored over the compensation of frequency-offsets Δf_mof signals from other signal sources m≠ m in order to better exploit the enhanced service HSDPA. To this end, J_m is chosen to J_m>1, whereas J_m=1 for m≠ m. Source importance weighting as explained above may be for instance accomplished in response to service indicators which are transmitted by the signal sources and which are translated in the receiver into service mode flags (e.g., a HSDPA-mode flag for a base station which offers HSDPA). Then, appropriate weighting values J_mare applied based on the activated flags in the receiver. As will be detailed further below, different weighting schemes may be used for J_m.
The second multiplier 212 multiplies the output of the first multiplier 211 by Δf_m. Thus, the weighting unit 210 generates at an output 213 of the second multiplier 212 values J_m·C_m·Δf_mand at an output 214 of the first multiplier 211 values J_m·C_m.
These outputs 213, 214 of the weighting unit 210 are input to the averaging unit 220. The averaging unit 220 comprises a first accumulator (AC) 221, a second accumulator (AC) 222 and a divider (DIV) 223. The accumulators 221, 222 accumulate their inputs over m=1, . . . , M and pass their respective results to the divider 223. The divider 223 generates the overall frequency-offset Δf according to
$\begin{matrix} \overline{Δ f} = \frac{\sum_{m = 1}^{M} J_{m} C_{m} Δ f_{m}}{\sum_{m = 1}^{M} J_{m} C_{m}} . & (13) \end{matrix}$
Averaging schemes different from the averaging scheme exemplified above may be employed.
The overall frequency-offset estimate Δf may be output periodically at a time interval given by L_avg. For instance, in an UMTS communications system, using as symbols r_i,mpilot symbols transmitted via the common pilot channel (CPICH), assuming L_avg=150 yields one value of Δf per frame (frame duration is 10 ms).
Further, with regard to circuitry design, as signal processing upstream of the averaging unit 220 is accomplished separately for each signal source m, it is either possible to duplicate the components arranged upstream of the averaging unit 220 or to operate them in a time multiplex cycle running successively over m=1, . . . , M.
In the example described above, both the frequency-offsets Δf_mand the quantities C_mrepresentative of a signal-to-noise ratio and used as weighting factors are computed on the basis of the same input data and on the basis of a differential phase estimation approach. That way, weighting factors C_minclude all effects relevant for weighting but not known by external sources and exclude all effects external sources are subject to while the frequency-offset estimates Δf_mare not. Further, as the weighting factors C_mmay be based on the same time interval as the frequency-offset estimates Δf_m, errors introduced by non-synchronized weighting factor sources are excluded. Thus, the above scheme for generating weighting factors C_mfrequency-offset estimates Δf_mmay be beneficial in view of low hardware implementation expenditure and high accuracy. However, instead of computing the frequency-offsets Δf_mfrom the averaged phasors P_m ^av, frequency-offsets Δf_mas used by the combiner 104 may also be computed elsewhere in the receiver, i.e. by different algorithms using other signal processing devices not shown in the foregoing. In this case, the argument unit 105 may be omitted.
According to another embodiment, the concept of source-importance weighting may be implemented in virtually any combiner independent of how frequency-offset estimates Δf_mand quantities representative of the signal-to-noise ratios are established. FIG. 4 illustrates a circuitry according to the general concept of source-importance weighting adjustment for an overall frequency-offset Δf calculation. Combiner 304 may be similar to combiner 104 as shown in FIG. 3. Optional discarding unit 200 may be omitted. Averaging unit 320 may be designed identical to averaging unit 220. Weighting unit 310 may, for example, use signal-to-noise ratios SNR_mor values C_mfor signal quality weighting purposes. Again, the signal source-specific frequency-offset estimates Δf_mmay be provided by different circuitry than shown in FIG. 2. Additionally, signal source-specific signal-to-noise ratios SNR_mcommunicated to the weighting unit 310 may be computed in a conventional fashion, i.e. not on the basis of a differential phase estimation approach (by using phasors) as explained above, but elsewhere in the receiver.
Moreover, it is possible in one embodiment that weighting of the frequency-offset estimates Δf_mwith the quantities representative of the signal-to-noise ratios is omitted. In this case, frequency-offset estimates Δf_mare simply weighted by the signal content dependent weighting factors J_m. Then, the combiner 104 may not receive the quantities C_mrepresentative of a signal-to-noise ratio. Thus, the discarding unit 200 and the multiplier 211 can be omitted and the signal content dependent weighting factors J_mmay be coupled to the input of the multiplier 212 (which receives at the other input the frequency-offset estimates Δf_m). As to FIG. 4, in this case the combiner 304 must not receive the quantities SNR_m(which may be represented by the quantities C_min FIG. 3).
Register 330 contains a number of signal source service flags each associated with a specific signal source (e.g., base station) m and indicative of whether the respective signal source m provides a specific service, for instance HSDPA. If a flag has the value 1, the service is provided, and if the flag has the value 0, the service is not provided. Here, signal source m (and probably also other signal sources) provides this service. If the flag has the value 1, multiplexer 340 is controlled to output J>1. Otherwise, if the flag has the value 0, multiplexer 340 outputs the value 1. Thus, if during a communication to multiple signal sources one signal source activates a specific service mode (e.g. HSDPA), the frequency-offset compensation in the receiver is improved for the signal transmitted by signal source m. It is apparent from the above that for different services different values of J may be used in order to differentiate between the importance of a plurality of services.
Features relating to the concept of source-importance weighting explained in conjunction with the circuitry shown in FIG. 4 may be applied in the aforementioned embodiments and vice versa. Further, features relating to the concept of estimating an overall frequency-offset Δf on the basis of a plurality of frequency-offset estimates Δf_mas presented in conjunction with FIGS. 1 to 3 are applicable to the circuitry as shown in FIG. 4 and vice versa.
According to another embodiment of the invention, circuitry for calculating a signal-to-noise ratio is shown in FIG. 5. Such circuitry may be used in virtually any kind of receiver in communications systems, i.e. its use is neither limited to mobile communications systems nor to radio systems. Concerning mobile communication systems, it may be implemented in a base station or in a mobile station or both. As known in the art, estimates of signal-to-noise ratio are used by various circuit parts in a receiver. For instance, in a RAKE receiver, outputs of single RAKE fingers can be combined in a combiner using signal-to-noise ratio estimates associated with each finger output (i.e. associated with a component of a signal which is transmitted via a specific propagation path of the radio channel). Thus, signal-to-noise ratio estimation according to this embodiment of the invention may, for example, also be used to provide signal-to-noise ratio estimates for a RAKE combiner.
The circuitry depicted in FIG. 5 comprises a phase-shift estimation unit 400 operative to generate averaged phasors p^avfrom an incoming complex-valued data stream r_i. The phase-shift estimation unit 400 comprises a phasor generation unit 100 and an averaging unit 101, which may be designed according to the corresponding units explained above. The averaged phasor p_avis thus formed according to equation (2) (without index m). As already mentioned, the averaged phasor p^avis representative of a phase-shift that occurred in the input data signal r_iover a number L_avgof consecutive data samples. L_avgmay be equal to 1 but typically, a greater averaging length is used.
The averaged phasors p^avare fed into a phase-shift processing unit 401. Phase-shift processing unit 401 comprises a first square (SQR) unit 402, a second square (SQR) unit 403, an adder 404, a root extraction unit 405 and a divider (DIV) 406. The first square unit 402 receives the real part of the averaged phasor p^avand its output is connected to a first input of the adder 404. The second square unit 403 receives the imaginary part of the averaged phasor p^avand its output is connected to a second input of the adder 404. At an output of the adder 404, quantity Q according to equation (3) is calculated.
In root extraction unit 405, Q^1/2is calculated. The square root of Q corresponds to the absolute value of the complex-valued averaged phasor p^avand is proportional to the signal-to-noise ratio SNR.
Provided that the energy level per data sample is Pt_s, the signal-to-noise ration SNR is to be calculated according to
$\begin{matrix} SNR = \frac{\sqrt{Q}}{2 {Pt}_{s} L_{avg}} . & (14) \end{matrix}$
To this end, the divider 406 divides the square root of Q by 2·Pt_s·L_avgand outputs one estimate of signal-to-noise ratio estimate SNR per L_avgdata samples r_i.
In many cases, it may not be necessary to calculate the signal-to-noise ratio but it may be sufficient to generate a quantity which is proportional to the signal-to-noise ratio SNR or even a quantity which is representative of the signal-to-noise ratio SNR according to a unique functional relationship. In such cases, the phase-shift processing unit 401 may e.g. output the result of the adder 404 (optionally divided by L_avg ²) or the output of the root extraction unit 405 (optionally divided by L_avg), respectively.
Although specific embodiments have been illustrated and described herein, it will be appreciated by those of ordinary skill in the art, that any arrangement which is calculated to achieve the same purpose may be substituted for the specific embodiments shown. It is to be understood, that the above description is intended to be illustrative and not restrictive. This application is intended to cover any adaptations or variations of the invention. Combinations of the above embodiments and many other embodiments will be apparent to those of skill in the art upon reading and understanding the above description. The scope of the invention includes any other embodiments and applications in which the above structures and methods may be used. The scope of the invention should, therefore, be determined with reference to the appended claims along with the scope of equivalents to which such claims are entitled.
It is emphasized that the Abstract is provided to comply with 37 C.F.R. section 1.72(b) requiring an abstract that will allow the reader to quickly ascertain the nature and gist of the technical disclosure. It is submitted with the understanding, that it will not be used to interpret or limit the scope or meaning of the claims.

Claims

1. A device for calculating an overall frequency-offset estimate from a plurality of frequency-offset estimates, comprising:

a receiver unit configured to receive a plurality of complex-valued data streams;

a phase shift estimation unit configured to generate a phase shift quantity representative of a phase shift over a number of consecutive data samples in the respective data stream;

a phase shift processing unit coupled to the phase shift estimation unit and configured to calculate a quantity related to the signal-to-noise ratio associated with a data stream on the basis of the respective phase shift quantity;

a frequency-offset estimation unit configured to generate frequency-offset estimates associated with the respective data streams; and

a combiner configured to calculate the overall frequency-offset estimate based on the frequency-offset estimates and the quantities related to the signal-to-noise ratios.

2. The device according to claim 1, wherein an input of the frequency-offset estimation unit is coupled to an output of the phase shift estimation unit.

3. The device according to claim 1, wherein the phase shift estimation unit comprises a multiplier for multiplying a data sample with a conjugate complex of a preceding data sample in the received data stream.

4. The device according to claim 3, wherein the phase shift estimation unit comprises an averaging unit having an input coupled to an output of the multiplier.

5. The device according to claim 1, wherein the combiner comprises a comparator configured to compare the quantities related to the signal-to-noise ratios with a threshold value.

6. The device according to claim 1, wherein the combiner comprises a first weighting stage configured to weight the quantities related to the signal-to-noise ratios with a first weight factor dependent on a signal content of the data stream associated with the respective signal-to-noise ratio.

7. The device according to claim 6, wherein the first weight factor is dependent on a service indicator transmitted by a signal source with respect to the respective data stream.

8. The device according to claim 1, wherein the combiner comprises a second weighting stage configured to weight the frequency-offset estimates with a second weight factor dependent on the signal-to-noise ratio associated with the respective frequency-offset estimate.

9. The device according to claim 8, wherein the combiner comprises an averaging stage configured to average the weighted frequency-offset estimates to generate the overall frequency-offset estimate.

10. A device for calculating an overall frequency-offset estimate from a plurality of frequency-offset estimates, comprising:

a weighting stage configured to weight the frequency-offset estimates with a respective weight factor that is dependent on a signal content of the signal from which the respective frequency-offset estimate was derived; and

an averaging stage configured to average the weighted frequency-offset estimates to generate the overall frequency-offset estimate.

11. The device of claim 10, further comprising:

a calculation unit configured to calculate a plurality of quantities which are each related to a signal-to-noise ratio associated with a signal from which the respective frequency-offset estimate was derived,

wherein the weighting stage is further configured to weight the frequency-offset estimates with the quantities which are each related to a signal-to-noise ratio.

12. The device of claim 10, wherein the weight factor is dependent on whether or not a specific service is supported by a signal source transmitting the respective signal.

13. A device for calculating a quantity related to a signal-to-noise ratio of a complex-valued data stream, comprising:

a phase shift estimation unit configured to generate a phase shift quantity representative of a phase shift occurred over a number of consecutive samples in the data stream, and

a phase shift processing unit coupled to the phase shift estimation unit and configured to calculate a quantity related to the signal-to-noise ratio associated with the complex-valued data stream on the basis of the respective phase shift quantity.

14. The device of claim 13, wherein the phase shift processing unit comprises a squaring element.

15. The device of claim 13, wherein the phase shift processing unit comprises a dividing stage.

16. The device of claim 15, wherein the divisor used in the dividing stage comprises a factor representative of the power level of the complex-valued data stream.

17. A method for calculating an overall frequency-offset estimate from a plurality of frequency-offset estimates, comprising:

receiving a plurality of complex-valued data streams;

estimating for each data stream a phase shift quantity representative of a phase shift occurring over a number of consecutive data samples in the respective data stream;

calculating for each data stream a quantity related to the signal-to-noise ratio in the respective data stream on the basis of the phase shift quantity associated with the respective data stream;

generating frequency-offset estimates associated with the respective data streams using the estimated phase shift quantities and the quantities related to the signal-to-noise ratio; and

combining the frequency-offset estimates on the basis of the quantities related to the signal-to-noise ratios to yield the overall frequency-offset estimate.

18. The method of claim 17, wherein estimating the phase shift quantity comprises multiplying a data sample by a complex conjugate of a preceding data sample in the respective data stream.

19. The method of claim 18, wherein estimating the phase shift quantity comprises summing a plurality of multiplication results.

20. The method of claim 17, wherein calculating the quantity related to the signal-to-noise ratio comprises squaring the respective phase shift quantity.

21. The method of claim 17, wherein combining the frequency-offset estimates comprises weighting the quantities related to the signal-to-noise ratios with a first weight factor that is dependent on a signal content of the data stream associated with the respective signal-to-noise ratio.

22. A method for calculating an overall frequency-offset estimate from a plurality of frequency-offset estimates, comprising:

weighting the frequency-offset estimates with a respective weight factor that is dependent on the signal content of the signal from which the respective frequency-offset estimate was derived; and

averaging the weighted frequency-offset estimates to generate the overall frequency-offset estimate.

23. The method of claim 22, further comprising:

calculating a plurality of quantities that are each related to a signal-to-noise ratio associated with a signal from which the respective frequency-offset estimate was derived; and

weighting the frequency-offset estimates with the quantities that are each related to a signal-to-noise ratio.

24. The method of claim 22, wherein the weight factor is adjusted dependent on whether or not a specific service is supported by a signal source transmitting the respective signal.

25. A method for calculating a quantity related to a signal-to-noise ratio of a complex-valued data stream, comprising:

generating a phase shift quantity representative of a phase shift that occurred over a number of consecutive samples in the data stream; and

calculating a quantity related to the signal-to-noise ratio associated with a the complex-valued data stream based on the respective phase shift quantity.

26. The method of claim 25, wherein calculating a quantity related to the signal-to-noise ratio comprises calculating the absolute value of the phase shift quantity.

27. The method of claim 26, wherein calculating a quantity related to the signal-to-noise ratio comprises dividing the calculated absolute value of the phase shift quantity by a factor representative of the power level of the complex-valued data stream.